prepare_tp_module# composer.distributed.prepare_tp_module(model, tp_config)[source]# Prepare a module (assumed ComposerModel) for use with tensor parallel.